Solar Simulator

[BradB]

https://react-whc5s3.stackblitz.io

I posted a Java version of this a few years back and got zero interest, but I do use it for my own tank a lot. It simulates sunlight at various depths and times in the ocean.

Some observations:

The actinic sunrise/sunset people use in their tanks is unnatural, and natural sunset/sunrise is mostly the same spectrum as high noon, with more reduction in reds and blues than yellows and greens.

I need to double check PAR calculations (3443 at high noon seems too high), but we need a lot more PAR than we think to simulate the real ocean at any depth. This does not necessarily translate into coral health.

Aquarium lighting is severely lacking in green and yellow light. This is not used as much by the coral and usually results in poorer coral coloration. It is part of why everything looks so much more vivid in our tanks than diving. People using natural sunlight over there tank should try filtering out most of the green and yellow, and a little red (people either filter all light equally or filter out longer wavelengths).

Assuming the calculations are correct, at the surface at noon, the ocean gets 136 watts/meter of light shorter in wavelength than the 440nm used by Royal Blue LEDs. It may not be valid to assume every coral and fish can live in surface UV, but it is a good starting point. For example, my tank is 1.48645 square meters, my UV LEDs are about 50% efficient, so I'd need 404 watts of violet/UV LEDs to duplicate surface radiation. I have nowhere near this amount right now. I haven't tried it, but would love to see a tank lit by 400nm-440nm LEDs only - completely unnatural, but should be safe for the inhabitants, grow coral and glow like crazy.

In the next few days, I will either verify or correct the PAR calculations. I am happy to share code and ideas with anyone interested.

 

[Maacc]

Very interesting if you can get the data collected. Are you doing this with field collected data or model extrapolation?
Either way it should be interesting.

 

[BradB]

Mostly model extrapolation. I did use a table for which wavelengths water absorbs, but the spectrum is made up assuming the sun is a 5778 degree black body, with the distance to earth and thickness of the atmosphere factored in.

 

[saltyfilmfolks]

Mostly model extrapolation. I did use a table for which wavelengths water absorbs, but the spectrum is made up assuming the sun is a 5778 degree black body, with the distance to earth and thickness of the atmosphere factored in.

I'd check that. 57 or 56 is the accepted blend of sun and sky. Fwiw sky is 10-20 alone. I'd guess sun as a black body is 25k or lower. It's a ball of fire. Candles are in the 18k range.

 

[Liquid360]

http://www.focustronic.net/

 

[BradB]

2 mistakes made porting from Java to the Javascript version have been fixed.

This gives a more reasonable 1738 PAR at high noon at the surface. With 676 PAR at 10 meters, we still need a lot more PAR than we think to simulate the real ocean.

At the surface at noon, the ocean gets 61 watts/meter of light shorter in wavelength than the 440nm used by Royal Blue LEDs. My 1.48645 square meter tank with 50% efficient UV Leds would need 178 watts of violet/UV LEDs to duplicate surface radiation.

 

[BradB]

I'd check that. 57 or 56 is the accepted blend of sun and sky. Fwiw sky is 10-20 alone. I'd guess sun as a black body is 25k or lower. It's a ball of fire. Candles are in the 18k range.

You are confusing 2 related concepts. Candles do not actually burn at 18,000 degrees Kelvin, anyone blowing out a birthday cake could never see their next birthday! When we say a light is 10K, what that originally meant was the light spectrum was similar to the light emitted from an actual 10,000K black body in a vacuum. This has changed somewhat - when we say a light is 10K, we really mean it is similar to the light emitted by other 10K lights. For 10K, this isn't too far off, but a 50K aquarium light is very blue, where a 50K black body is mostly harmful UV.

The sky is said to be 10k-20k, which is considerably bluer than the sun's light in a vacuum, which is correct and included in the model (Rayleigh scattering).

 

[saltyfilmfolks]

You are confusing 2 related concepts. Candles do not actually burn at 18,000 degrees Kelvin, anyone blowing out a birthday cake could never see their next birthday! When we say a light is 10K, what that originally meant was the light spectrum was similar to the light emitted from an actual 10,000K black body in a vacuum. This has changed somewhat - when we say a light is 10K, we really mean it is similar to the light emitted by other 10K lights. For 10K, this isn't too far off, but a 50K aquarium light is very blue, where a 50K black body is mostly harmful UV.

The sky is said to be 10k-20k, which is considerably bluer than the sun's light in a vacuum, which is correct and included in the model (Rayleigh scattering).

It was a typo. Sorry.
Candles run in the 1800 range.
I have a color meter.
And I disagree , we do see 18,000 k regularly.

 

[BradB]

It was a typo. Sorry.
Candles run in the 1800 range.
I have a color meter.
And I disagree , we do see 18,000 k regularly.

You might see 18,000 k on your color meter, but you won't on a thermometer. This will be my last post on CCT vs blackbody radiation. Most introductory physics textbooks can explain this better than me.

 

[saltyfilmfolks]

You might see 18,000 k on your color meter, but you won't on a thermometer. This will be my last post on CCT vs blackbody radiation. Most introductory physics textbooks can explain this better than me.

I'm sorry. We're talking about basic color temperature correct?

 

[Liquid360]

Gotta love Google

 

[saltyfilmfolks]

Gotta love Google

And 30 years of color temping lights.

I dunno.

And those xenins only get to 5k when the bulb goes bad.

 

[A. grandis]

Very interesting if you can get the data collected. Are you doing this with field collected data or model extrapolation?
Either way it should be interesting.

Your avatar is very nice!! Please give the dog his or her favorite treat and tell it's from uncle Grandis!!
Grandis.

 

[Dennis Cartier]

In my old tank, I wrote up a python program to collect real lighting data from the weather Underground site using their API. My plan was to "replay" the calendar shifted to control my DIY led fixture PAR levels. I never got around to actually using the data that I collected up unfortunately. I still think modeling light output based on data from actual tropical locations would be an interesting project.

Dennis

 

[Liquid360]

In my old tank, I wrote up a python program to collect real lighting data from the weather Underground site using their API. My plan was to "replay" the calendar shifted to control my DIY led fixture PAR levels. I never got around to actually using the data that I collected up unfortunately. I still think modeling light output based on data from actual tropical locations would be an interesting project.

Dennis

Adjusting of course for global warming [emoji6]

 

[BradB]

In my old tank, I wrote up a python program to collect real lighting data from the weather Underground site using their API. My plan was to "replay" the calendar shifted to control my DIY led fixture PAR levels. I never got around to actually using the data that I collected up unfortunately. I still think modeling light output based on data from actual tropical locations would be an interesting project.
Dennis

Do you actually have this data still? I'd be interested in seeing it and knowing more about it. I have to assume it is all surface level.

It is interesting to build a tank to simulate natural conditions, but I don't know if it is worth it, beyond the fun of actually building it. The angle light comes from varies by 180 degrees on a real reef, and this would be a lot of work and expense. As I said, our reef tank spectrums look much better than real sunlight, and real sunlight is brighter than what we need. But knowing what the natural parameters are is worthwhile IMO.

 

[BradB]

And 30 years of color temping lights.

I dunno.

And those xenins only get to 5k when the bulb goes bad.

And after 30 years of gapping spark plugs I know it is 1.2mm but you say it weighs 6 grams? You aren't measuring wrong, you are referring to the wrong measurement. Both are degrees, so it can be confusing, but if you had to cut a 90 degree angle, would you use a color meter or a thermometer? 'Degree' can mean different things.

The sun is actually 5778 degrees Kelvin measured by a thermometer (obviously you can't go there and use a mercury thermometer, but the same thing is being measured as if you could). If you launched a perfect color meter into outer space, it would also measure sunlight at 5778 degrees Kelvin. Now the other light sources mentioned, candles, Xenon bulbs, metal halides, the sky, etc - they don't have much correlation between physical temperature and light spectrum. The sky might be 3 times the color temperature of the sun, but our sky is NOT 3 times the physical temperature of the sun's heat.

On Earth, the angle of the sun and our atmosphere affect the spectrum - typically the sky appears blue especially near noon, the sun appears more yellow/red especially at low angles and UV and shorter wave radiation that would kill us gets filtered out. A color meter on earth will may reasonably give various readings between 3,000 and 20,000 depending on the angle of the sun, and how much light from the sky is included.

 

[saltyfilmfolks]

And after 30 years of gapping spark plugs I know it is 1.2mm but you say it weighs 6 grams? You aren't measuring wrong, you are referring to the wrong measurement. Both are degrees, so it can be confusing, but if you had to cut a 90 degree angle, would you use a color meter or a thermometer? 'Degree' can mean different things.

The sun is actually 5778 degrees Kelvin measured by a thermometer (obviously you can't go there and use a mercury thermometer, but the same thing is being measured as if you could). If you launched a perfect color meter into outer space, it would also measure sunlight at 5778 degrees Kelvin. Now the other light sources mentioned, candles, Xenon bulbs, metal halides, the sky, etc - they don't have much correlation between physical temperature and light spectrum. The sky might be 3 times the color temperature of the sun, but our sky is NOT 3 times the physical temperature of the sun's heat.

On Earth, the angle of the sun and our atmosphere affect the spectrum - typically the sky appears blue especially near noon, the sun appears more yellow/red especially at low angles and UV and shorter wave radiation that would kill us gets filtered out. A color meter on earth will may reasonably give various readings between 3,000 and 20,000 depending on the angle of the sun, and how much light from the sky is included.

Got it. Thanks.

 

[Dennis Cartier]

Do you actually have this data still? I'd be interested in seeing it and knowing more about it. I have to assume it is all surface level.

It is interesting to build a tank to simulate natural conditions, but I don't know if it is worth it, beyond the fun of actually building it. The angle light comes from varies by 180 degrees on a real reef, and this would be a lot of work and expense. As I said, our reef tank spectrums look much better than real sunlight, and real sunlight is brighter than what we need. But knowing what the natural parameters are is worthwhile IMO.

I might have it on my old PC. If I recall, the API that they have allows you to pull a JSON feed for a 24 hour period. I was using a weather station in Grand Cayman. The hardest part was that they had a throttle on the API, so collecting up a years worth, was going to take quite awhile. Finding a weather station that pushed data on a constant and consistent basis was also a bit of a challenge. Probably more so now that the Caribbean has been savaged by recent hurricanes.

If I think of it this weekend, I will have a look for it.

Dennis

 

[saltyfilmfolks]

Silly me. I was color temping lights in my house and the sun from the surface of the planet where I live.